Acknowledgements

I am most indebted to Barney Pell for introducing me to Reinforcement Learning. Thanks to my supervisor John Daugman for his consistent support and advice. Thanks to Richard Prager, Tony Prescott and a number of anonymous referees for much essential criticism.

Credit to Ian Lewis and Gavin Rummery for important contributions, and thanks also to Ralph Becket, Marcel Hernandez, Steve Hodges, Clare Jackson, Dave Palfrey, Tony Robinson, Michael K.Sahota and Chen K.Tham for useful comments on this work. There are many other people that I will not list, with whom I have had useful and inspiring conversations about my work and related topics. They have made me so glad I chose to come to Cambridge.

Thanks to John Murphy for encouraging me to do a PhD. Sincere thanks to William Clocksin and Roger Needham for getting me started. This work was supported mainly by the British Council and by the University of Cambridge Computer Laboratory. I am grateful to the U.S. Office of Naval Research and the M.R. Bauer Foundation for a student travelship. I am also grateful for contributions from Trinity Hall, Cambridge, from the Cambridge Philosophical Society and from the University of Cambridge Board of Graduate Studies.

Thanks to Frank for his comments, thanks to Richard for astonishingly reading the whole thesis, and thanks in general to Richard, Frank and Joe for many years of entertainment. Much love to Elizabeth, and thanks for being there for the good part.

Dedication

This dissertation is dedicated to my mother and father,
to whom I owe everything.

In loving memory of my grandmother.

Appendix A

My PhD "family tree" (Who supervised who)

Return to Contents page.